[Rd] R-1.7.0-beta: in check-all test-IsAs fails on WinNT4.0

[Uwe Ligges]

Uwe Ligges wrote:
> Todays R-1.7.0-beta (see below) compiles fine on, but it doesn't pass
>  make check-all
> on WinNT4.0, SP6.
> 
> Looking closer, test-IsAs fails:
> 
> D:\uwe-ns\R-1.7.0\tests>make -f Makefile.win test-IsAs
> running tests of consistency of as/is.*
> make[1]: Entering directory `/cygdrive/d/uwe-ns/R-1.7.0/tests'
> creating 'isas-tests.R'
> make[1]: *** [isas-tests.R] Error 253
> make[1]: Leaving directory `/cygdrive/d/uwe-ns/R-1.7.0/tests'
> make: *** [test-IsAs] Error 2
> 
> Rterm crashes with a "stack overflow".
> 
> The strange thing: I can run the test manually without any problems (I 
> tried to invoke rterm the same way as in the makefiles -- well, 
> apparently I just *tried*)!


Yep. I did it again (and hopefully correct) and in the following case 
Rterm.exe crashes with a stack overflow:

D:\uwe-ns\R-1.7.0\tests>d:\uwe-ns\r-1.7.0\bin\rterm --vanilla --slave 
LC_ALL=C SRCDIR=. < isas-tests.Rin

which is the way R is called in the Makefile for test-IsAs, AFAICS.

I know, it's much too late for this report, but in fact I was too busy 
with an S course during the last 7 days to look closer. Sorry.

Uwe Ligges


> It looks like the file "isas-tests.R" which should be sink()ed in 
> "isas-tests.Rin" won't be created. Any ideas?
> 
> Uwe Ligges
> 
>          _
> platform i386-pc-mingw32
> arch     i386
> os       mingw32
> system   i386, mingw32
> status   Under development (unstable)
> major    1
> minor    7.0
> year     2003
> month    04
> day      12
> language R
> 
> ______________________________________________
> R-devel at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel